Motion estimation and segmentation of cardiac magnetic resonance images using variational and level set techniques

A new method for motion estimation and segmentation of cardiac magnetic resonance images, based on variational and level set techniques is presented. The variational method for motion estimation is based on a total variation approach following a matrix implementation. It has been integrated with a level set method for segmenting the endocardial wall, providing a valuable tool for motion estimation and segmentation together. The results obtained have been compared with a reference method. The robustness and accuracy of these results have been proved.